/external/llvm-project/llvm/test/CodeGen/X86/ |
D | fp-roundeven.ll | 2 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse2 | FileCheck %s --check-prefix=SSE2 8 ; SSE2-LABEL: roundeven_f32: 9 ; SSE2: ## %bb.0: 10 ; SSE2-NEXT: jmp _roundevenf ## TAILCALL 26 ; SSE2-LABEL: roundeven_f64: 27 ; SSE2: ## %bb.0: 28 ; SSE2-NEXT: jmp _roundeven ## TAILCALL 44 ; SSE2-LABEL: roundeven_v4f32: 45 ; SSE2: ## %bb.0: 46 ; SSE2-NEXT: subq $56, %rsp [all …]
|
D | ftrunc.ll | 2 …: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 7 ; SSE2-LABEL: trunc_unsigned_f32: 8 ; SSE2: # %bb.0: 9 ; SSE2-NEXT: cvttss2si %xmm0, %rax 10 ; SSE2-NEXT: movl %eax, %eax 11 ; SSE2-NEXT: xorps %xmm0, %xmm0 12 ; SSE2-NEXT: cvtsi2ss %rax, %xmm0 13 ; SSE2-NEXT: retq 30 ; SSE2-LABEL: trunc_unsigned_f64: 31 ; SSE2: # %bb.0: [all …]
|
D | fp-round.ll | 2 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse2 | FileCheck %s --check-prefix=SSE2 8 ; SSE2-LABEL: round_f32: 9 ; SSE2: ## %bb.0: 10 ; SSE2-NEXT: jmp _roundf ## TAILCALL 45 ; SSE2-LABEL: round_f64: 46 ; SSE2: ## %bb.0: 47 ; SSE2-NEXT: jmp _round ## TAILCALL 73 ; SSE2-LABEL: round_v4f32: 74 ; SSE2: ## %bb.0: 75 ; SSE2-NEXT: subq $56, %rsp [all …]
|
D | funnel-shift.ll | 2 ; RUN: llc < %s -mtriple=i686-- -mattr=sse2 | FileCheck %s --check-prefixes=ANY,X32-SSE2 20 ; X32-SSE2-LABEL: fshl_i32: 21 ; X32-SSE2: # %bb.0: 22 ; X32-SSE2-NEXT: movb {{[0-9]+}}(%esp), %cl 23 ; X32-SSE2-NEXT: movl {{[0-9]+}}(%esp), %edx 24 ; X32-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax 25 ; X32-SSE2-NEXT: shldl %cl, %edx, %eax 26 ; X32-SSE2-NEXT: retl 42 ; X32-SSE2-LABEL: fshl_i37: 43 ; X32-SSE2: # %bb.0: [all …]
|
D | vector-ext-logic.ll | 2 ; RUN: llc < %s -mtriple=x86_64-- -mattr=+sse2 | FileCheck %s --check-prefix=SSE2 6 ; SSE2-LABEL: zext_and_v8i32: 7 ; SSE2: # %bb.0: 8 ; SSE2-NEXT: movdqa %xmm0, %xmm2 9 ; SSE2-NEXT: pand %xmm1, %xmm2 10 ; SSE2-NEXT: pxor %xmm1, %xmm1 11 ; SSE2-NEXT: movdqa %xmm2, %xmm0 12 ; SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[3],x… 13 ; SSE2-NEXT: punpckhwd {{.*#+}} xmm2 = xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[7],x… 14 ; SSE2-NEXT: movdqa %xmm2, %xmm1 [all …]
|
D | vselect-minmax.ll | 2 …RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 10 ; SSE2-LABEL: test1: 11 ; SSE2: # %bb.0: # %entry 12 ; SSE2-NEXT: movdqa %xmm1, %xmm2 13 ; SSE2-NEXT: pcmpgtb %xmm0, %xmm2 14 ; SSE2-NEXT: pand %xmm2, %xmm0 15 ; SSE2-NEXT: pandn %xmm1, %xmm2 16 ; SSE2-NEXT: por %xmm2, %xmm0 17 ; SSE2-NEXT: retq 35 ; SSE2-LABEL: test2: [all …]
|
D | pr34080.ll | 2 ; RUN: llc < %s -mtriple=x86_64-apple-darwin -mattr=+sse2 | FileCheck %s --check-prefix=SSE2 3 … -mtriple=x86_64-apple-darwin -mattr=+sse2 -mcpu=x86-64 | FileCheck %s --check-prefix=SSE2-SCHEDULE 10 ; SSE2-LABEL: _Z1fe: 11 ; SSE2: ## %bb.0: ## %entry 12 ; SSE2-NEXT: pushq %rbp 13 ; SSE2-NEXT: .cfi_def_cfa_offset 16 14 ; SSE2-NEXT: .cfi_offset %rbp, -16 15 ; SSE2-NEXT: movq %rsp, %rbp 16 ; SSE2-NEXT: .cfi_def_cfa_register %rbp 17 ; SSE2-NEXT: fldt 16(%rbp) [all …]
|
D | vector_splat-const-shift-of-constmasked.ll | 2 …: llc < %s -mtriple=i686-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X86-SSE2 5 …: llc < %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=X64-SSE2 18 ; X86-SSE2-LABEL: test_128_i8_x_16_7_mask_lshr_1: 19 ; X86-SSE2: # %bb.0: 20 ; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0 21 ; X86-SSE2-NEXT: psrlw $1, %xmm0 22 ; X86-SSE2-NEXT: pand {{\.LCPI.*}}, %xmm0 23 ; X86-SSE2-NEXT: retl 32 ; X64-SSE2-LABEL: test_128_i8_x_16_7_mask_lshr_1: 33 ; X64-SSE2: # %bb.0: [all …]
|
D | vec_setcc-2.ll | 2 … llc < %s -o - -mtriple=x86_64-apple-darwin -mattr=+sse2 | FileCheck %s --check-prefixes=CHECK,SSE2 8 ; SSE2-LABEL: loop_no_const_reload: 9 ; SSE2: ## %bb.0: ## %entry 10 ; SSE2-NEXT: testl %edx, %edx 11 ; SSE2-NEXT: je LBB0_3 12 ; SSE2-NEXT: ## %bb.1: ## %for.body.preheader 13 ; SSE2-NEXT: xorl %eax, %eax 14 ; SSE2-NEXT: movdqa {{.*#+}} xmm0 = [25,25,25,25,25,25,25,25] 15 ; SSE2-NEXT: pxor %xmm1, %xmm1 16 ; SSE2-NEXT: .p2align 4, 0x90 [all …]
|
D | vector-reduce-smax.ll | 2 …RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 15 ; SSE2-LABEL: test_v2i64: 16 ; SSE2: # %bb.0: 17 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3] 18 ; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648] 19 ; SSE2-NEXT: movdqa %xmm0, %xmm3 20 ; SSE2-NEXT: pxor %xmm2, %xmm3 21 ; SSE2-NEXT: pxor %xmm1, %xmm2 22 ; SSE2-NEXT: movdqa %xmm3, %xmm4 23 ; SSE2-NEXT: pcmpgtd %xmm2, %xmm4 [all …]
|
D | vector-reduce-smin.ll | 2 …RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 15 ; SSE2-LABEL: test_v2i64: 16 ; SSE2: # %bb.0: 17 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3] 18 ; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [2147483648,2147483648] 19 ; SSE2-NEXT: movdqa %xmm0, %xmm3 20 ; SSE2-NEXT: pxor %xmm2, %xmm3 21 ; SSE2-NEXT: pxor %xmm1, %xmm2 22 ; SSE2-NEXT: movdqa %xmm2, %xmm4 23 ; SSE2-NEXT: pcmpgtd %xmm3, %xmm4 [all …]
|
D | vector-reduce-umin.ll | 2 …RUN: llc < %s -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefixes=SSE,SSE2 15 ; SSE2-LABEL: test_v2i64: 16 ; SSE2: # %bb.0: 17 ; SSE2-NEXT: pshufd {{.*#+}} xmm1 = xmm0[2,3,2,3] 18 ; SSE2-NEXT: movdqa {{.*#+}} xmm2 = [9223372039002259456,9223372039002259456] 19 ; SSE2-NEXT: movdqa %xmm0, %xmm3 20 ; SSE2-NEXT: pxor %xmm2, %xmm3 21 ; SSE2-NEXT: pxor %xmm1, %xmm2 22 ; SSE2-NEXT: movdqa %xmm2, %xmm4 23 ; SSE2-NEXT: pcmpgtd %xmm3, %xmm4 [all …]
|
D | widen_conv-3.ll | 2 ; RUN: llc < %s -mtriple=i686-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86-SSE2 4 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64-SSE2 10 ; X86-SSE2-LABEL: convert_v2i16_to_v2f32: 11 ; X86-SSE2: # %bb.0: # %entry 12 ; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax 13 ; X86-SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0,0,1,1,2,2,3,3] 14 ; X86-SSE2-NEXT: psrad $16, %xmm0 15 ; X86-SSE2-NEXT: cvtdq2ps %xmm0, %xmm0 16 ; X86-SSE2-NEXT: movlps %xmm0, (%eax) 17 ; X86-SSE2-NEXT: retl [all …]
|
D | funnel-shift-rot.ll | 2 ; RUN: llc < %s -mtriple=i686-- -mattr=sse2 | FileCheck %s --check-prefixes=ANY,X32-SSE2 20 ; X32-SSE2-LABEL: rotl_i8_const_shift: 21 ; X32-SSE2: # %bb.0: 22 ; X32-SSE2-NEXT: movb {{[0-9]+}}(%esp), %al 23 ; X32-SSE2-NEXT: rolb $3, %al 24 ; X32-SSE2-NEXT: retl 37 ; X32-SSE2-LABEL: rotl_i8_const_shift1: 38 ; X32-SSE2: # %bb.0: 39 ; X32-SSE2-NEXT: movb {{[0-9]+}}(%esp), %al 40 ; X32-SSE2-NEXT: rolb %al [all …]
|
D | widen_conv-4.ll | 2 ; RUN: llc < %s -mtriple=i686-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86-SSE2 4 ; RUN: llc < %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64-SSE2 10 ; X86-SSE2-LABEL: convert_v7i16_v7f32: 11 ; X86-SSE2: # %bb.0: # %entry 12 ; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax 13 ; X86-SSE2-NEXT: pxor %xmm1, %xmm1 14 ; X86-SSE2-NEXT: movdqa %xmm0, %xmm2 15 ; X86-SSE2-NEXT: punpckhwd {{.*#+}} xmm2 = xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[… 16 ; X86-SSE2-NEXT: cvtdq2ps %xmm2, %xmm2 17 ; X86-SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[… [all …]
|
/external/llvm/test/Analysis/CostModel/X86/ |
D | testshiftshl.ll | 1 ; RUN: llc -mtriple=x86_64-apple-darwin -mcpu=core2 < %s | FileCheck --check-prefix=SSE2-CODEGEN %s 2 …triple=x86_64-apple-darwin -mcpu=core2 -cost-model -analyze < %s | FileCheck --check-prefix=SSE2 %s 7 ; SSE2: shift2i16 8 ; SSE2: cost of 4 {{.*}} shl 9 ; SSE2-CODEGEN: shift2i16 10 ; SSE2-CODEGEN: psllq 19 ; SSE2: shift4i16 20 ; SSE2: cost of 10 {{.*}} shl 21 ; SSE2-CODEGEN: shift4i16 22 ; SSE2-CODEGEN: pmuludq [all …]
|
D | testshiftlshr.ll | 1 ; RUN: llc -mtriple=x86_64-apple-darwin -mcpu=core2 < %s | FileCheck --check-prefix=SSE2-CODEGEN %s 2 …triple=x86_64-apple-darwin -mcpu=core2 -cost-model -analyze < %s | FileCheck --check-prefix=SSE2 %s 7 ; SSE2: shift2i16 8 ; SSE2: cost of 4 {{.*}} lshr 9 ; SSE2-CODEGEN: shift2i16 10 ; SSE2-CODEGEN: psrlq 19 ; SSE2: shift4i16 20 ; SSE2: cost of 16 {{.*}} lshr 21 ; SSE2-CODEGEN: shift4i16 22 ; SSE2-CODEGEN: psrld [all …]
|
D | testshiftashr.ll | 1 ; RUN: llc -mtriple=x86_64-apple-darwin -mcpu=core2 < %s | FileCheck --check-prefix=SSE2-CODEGEN %s 2 …triple=x86_64-apple-darwin -mcpu=core2 -cost-model -analyze < %s | FileCheck --check-prefix=SSE2 %s 7 ; SSE2: shift2i16 8 ; SSE2: cost of 12 {{.*}} ashr 9 ; SSE2-CODEGEN: shift2i16 10 ; SSE2-CODEGEN: psrlq 19 ; SSE2: shift4i16 20 ; SSE2: cost of 16 {{.*}} ashr 21 ; SSE2-CODEGEN: shift4i16 22 ; SSE2-CODEGEN: psrad [all …]
|
/external/llvm-project/llvm/test/Analysis/CostModel/X86/ |
D | testshiftlshr.ll | 1 ; RUN: llc -mtriple=x86_64-apple-darwin -mcpu=core2 < %s | FileCheck --check-prefix=SSE2-CODEGEN %s 2 …triple=x86_64-apple-darwin -mcpu=core2 -cost-model -analyze < %s | FileCheck --check-prefix=SSE2 %s 7 ; SSE2-LABEL: shift2i16 8 ; SSE2: cost of 32 {{.*}} lshr 9 ; SSE2-CODEGEN-LABEL: shift2i16 10 ; SSE2-CODEGEN: psrlw 19 ; SSE2-LABEL: shift4i16 20 ; SSE2: cost of 32 {{.*}} lshr 21 ; SSE2-CODEGEN-LABEL: shift4i16 22 ; SSE2-CODEGEN: psrlw [all …]
|
D | testshiftashr.ll | 1 ; RUN: llc -mtriple=x86_64-apple-darwin -mcpu=core2 < %s | FileCheck --check-prefix=SSE2-CODEGEN %s 2 …triple=x86_64-apple-darwin -mcpu=core2 -cost-model -analyze < %s | FileCheck --check-prefix=SSE2 %s 7 ; SSE2-LABEL: shift2i16 8 ; SSE2: cost of 32 {{.*}} ashr 9 ; SSE2-CODEGEN-LABEL: shift2i16 10 ; SSE2-CODEGEN: psraw 19 ; SSE2-LABEL: shift4i16 20 ; SSE2: cost of 32 {{.*}} ashr 21 ; SSE2-CODEGEN-LABEL: shift4i16 22 ; SSE2-CODEGEN: psraw [all …]
|
D | testshiftshl.ll | 1 ; RUN: llc -mtriple=x86_64-apple-darwin -mcpu=core2 < %s | FileCheck --check-prefix=SSE2-CODEGEN %s 2 …triple=x86_64-apple-darwin -mcpu=core2 -cost-model -analyze < %s | FileCheck --check-prefix=SSE2 %s 7 ; SSE2-LABEL: shift2i16 8 ; SSE2: cost of 32 {{.*}} shl 9 ; SSE2-CODEGEN-LABEL: shift2i16 10 ; SSE2-CODEGEN: pmullw 19 ; SSE2-LABEL: shift4i16 20 ; SSE2: cost of 32 {{.*}} shl 21 ; SSE2-CODEGEN-LABEL: shift4i16 22 ; SSE2-CODEGEN: pmullw [all …]
|
/external/llvm/test/CodeGen/X86/ |
D | vselect-minmax.ll | 2 … -mtriple=x86_64-unknown-unknown -mattr=+sse2 | FileCheck %s --check-prefix=SSE --check-prefix=SSE2 10 ; SSE2-LABEL: test1: 11 ; SSE2: # BB#0: # %entry 12 ; SSE2-NEXT: movdqa %xmm1, %xmm2 13 ; SSE2-NEXT: pcmpgtb %xmm0, %xmm2 14 ; SSE2-NEXT: pand %xmm2, %xmm0 15 ; SSE2-NEXT: pandn %xmm1, %xmm2 16 ; SSE2-NEXT: por %xmm2, %xmm0 17 ; SSE2-NEXT: retq 35 ; SSE2-LABEL: test2: [all …]
|
D | widen_conv-4.ll | 2 …c < %s -mtriple=i686-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86 --check-prefix=X86-SSE2 4 …< %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64 --check-prefix=X64-SSE2 10 ; X86-SSE2-LABEL: convert_v7i16_v7f32: 11 ; X86-SSE2: # BB#0: # %entry 12 ; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax 13 ; X86-SSE2-NEXT: pxor %xmm1, %xmm1 14 ; X86-SSE2-NEXT: movdqa %xmm0, %xmm2 15 ; X86-SSE2-NEXT: punpckhwd {{.*#+}} xmm2 = xmm2[4],xmm1[4],xmm2[5],xmm1[5],xmm2[6],xmm1[6],xmm2[… 16 ; X86-SSE2-NEXT: cvtdq2ps %xmm2, %xmm2 17 ; X86-SSE2-NEXT: punpcklwd {{.*#+}} xmm0 = xmm0[0],xmm1[0],xmm0[1],xmm1[1],xmm0[2],xmm1[2],xmm0[… [all …]
|
D | widen_conv-3.ll | 2 …c < %s -mtriple=i686-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X86 --check-prefix=X86-SSE2 4 …< %s -mtriple=x86_64-unknown -mattr=+sse2 | FileCheck %s --check-prefix=X64 --check-prefix=X64-SSE2 10 ; X86-SSE2-LABEL: convert_v2i16_to_v2f32: 11 ; X86-SSE2: # BB#0: # %entry 12 ; X86-SSE2-NEXT: movl {{[0-9]+}}(%esp), %eax 13 ; X86-SSE2-NEXT: psllq $48, %xmm0 14 ; X86-SSE2-NEXT: psrad $16, %xmm0 15 ; X86-SSE2-NEXT: pshufd {{.*#+}} xmm0 = xmm0[1,3,2,3] 16 ; X86-SSE2-NEXT: cvtdq2ps %xmm0, %xmm0 17 ; X86-SSE2-NEXT: movss %xmm0, (%eax) [all …]
|
D | vector-bitreverse.ll | 2 …known-unknown -mattr=+sse2 | FileCheck %s --check-prefix=ALL --check-prefix=SSE --check-prefix=SSE2 1022 ; SSE2-LABEL: test_bitreverse_v16i8: 1023 ; SSE2: # BB#0: 1024 ; SSE2-NEXT: movdqa %xmm0, %xmm2 1025 ; SSE2-NEXT: psrlw $7, %xmm2 1026 ; SSE2-NEXT: movdqa {{.*#+}} xmm1 = [1,1,1,1,1,1,1,1,1,1,1,1,1,1,1,1] 1027 ; SSE2-NEXT: pand %xmm1, %xmm1 1028 ; SSE2-NEXT: pand %xmm2, %xmm1 1029 ; SSE2-NEXT: movdqa %xmm0, %xmm2 1030 ; SSE2-NEXT: psllw $7, %xmm2 [all …]
|